Artificial intelligence in marketing

By March 14, 2019Data
artificial intelligence?

These aren’t the robots you’re looking for. (Photo by Craig Sybert on Unsplash)

Since the term was coined in the 1950s, “artificial intelligence” (AI) has often brought to mind robots that suddenly become sentient and try to take over the world. (Robots that become sentient and give everyone a puppy and a hug must not make for compelling science fiction.)

In reality, AI is more benign than that—at least so far.

By simple definition, AI refers to a machine or system that can “learn” on its own. That means figuring things out in real time, not just completing pre-determined actions.

Within machine learning, there are two types of tasks: supervised and unsupervised. An example of the latter is the autonomous car. Self-driving vehicles don’t have a rigid set of instructions to follow to get from A to B. They have to instantly process and respond to changing conditions in order to function and safely arrive at their destination. Unsupervised learning like this requires the machine to infer relationships between data points without outside help.

Supervised learning, on the other hand, is much more commonly used, including by data-driven marketing companies like ours. We can use it to make predictions about the behavior of our clients’ customers based on the variables we have data on (i.e., demographics, spending patterns, personal interests, etc.). This can be incredibly helpful when trying to create an email that will not only get read but also move someone to action. We need to know whom we’re talking to in order to understand how to talk to them.

Artificial intelligence is a great tool for squeezing more actionable insight out of customer data. And there’s certainly plenty of information to work with. Every day, more than 2.5 quintillion bytes of data are created.

Our data team is currently exploring using artificial intelligence to:

  • Identify who among our target audiences are most inclined to open emails based on demographics and their relationship with the sender
  • Conduct sentiment analysis of email responses and survey or form submissions to gauge and plot attitudes over time
  • Detect which emails are most likely to be opened by certain customers based on content in the subject line
  • Categorize customer types and identify potential micro-segments based on previously unrecognized patterns (as in, each year the same 20 people book rooms for the same period of time at the same hotel–ends up, it’s for their family reunion)

Interesting stuff, and we’re in the early days of AI. There’s still hope for a puppy-gifting robot in our future. Or at least more intelligent, more effective, more human marketing.

(For more about AI, check out this article from Forbes.)